Stefan Mayer
@stefanmayer.bsky.social
Behavioral Science meets Data Science.
Assistant Professor of Marketing Analytics, University of Tübingen.
Aesthetics | Machine Learning | Open Science.
https://uni-tuebingen.de/en/148617
Assistant Professor of Marketing Analytics, University of Tübingen.
Aesthetics | Machine Learning | Open Science.
https://uni-tuebingen.de/en/148617
Reposted by Stefan Mayer
🚨 SynthNet is out 🚨
Researchers propose new constructs and measures faster than anyone can track. We (@anniria.bsky.social @ruben.the100.ci) built a search engine to check what already exists and help identify redundancies; indexing 74,000 scales from ~31,500 instruments in APA PsycTests. 🧵1/3
Researchers propose new constructs and measures faster than anyone can track. We (@anniria.bsky.social @ruben.the100.ci) built a search engine to check what already exists and help identify redundancies; indexing 74,000 scales from ~31,500 instruments in APA PsycTests. 🧵1/3
November 26, 2025 at 11:42 AM
🚨 SynthNet is out 🚨
Researchers propose new constructs and measures faster than anyone can track. We (@anniria.bsky.social @ruben.the100.ci) built a search engine to check what already exists and help identify redundancies; indexing 74,000 scales from ~31,500 instruments in APA PsycTests. 🧵1/3
Researchers propose new constructs and measures faster than anyone can track. We (@anniria.bsky.social @ruben.the100.ci) built a search engine to check what already exists and help identify redundancies; indexing 74,000 scales from ~31,500 instruments in APA PsycTests. 🧵1/3
Reposted by Stefan Mayer
1. LLM-generated code tries to run code from online software packages. Which is normal but
2. The packages don’t exist. Which would normally cause an error but
3. Nefarious people have made malware under the package names that LLMs make up most often. So
4. Now the LLM code points to malware.
2. The packages don’t exist. Which would normally cause an error but
3. Nefarious people have made malware under the package names that LLMs make up most often. So
4. Now the LLM code points to malware.
LLMs hallucinating nonexistent software packages with plausible names leads to a new malware vulnerability: "slopsquatting."
LLMs can't stop making up software dependencies and sabotaging everything
: Hallucinated package names fuel 'slopsquatting'
www.theregister.com
April 12, 2025 at 11:43 PM
1. LLM-generated code tries to run code from online software packages. Which is normal but
2. The packages don’t exist. Which would normally cause an error but
3. Nefarious people have made malware under the package names that LLMs make up most often. So
4. Now the LLM code points to malware.
2. The packages don’t exist. Which would normally cause an error but
3. Nefarious people have made malware under the package names that LLMs make up most often. So
4. Now the LLM code points to malware.
Reposted by Stefan Mayer
I’ve long used FiveThirtyEight’s interactive “Hack Your Way To Scientific Glory” to illustrate the idea of p-hacking when I teach statistics. But ABC/Disney killed the site earlier this month :(
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
March 20, 2025 at 6:30 PM
I’ve long used FiveThirtyEight’s interactive “Hack Your Way To Scientific Glory” to illustrate the idea of p-hacking when I teach statistics. But ABC/Disney killed the site earlier this month :(
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
Reposted by Stefan Mayer
The problem with #rstats is portability. Shit like `df$x` has to be converted to `df€x` in Europe and `df£x` in the UK but nobody talks about this.
March 14, 2025 at 12:42 PM
The problem with #rstats is portability. Shit like `df$x` has to be converted to `df€x` in Europe and `df£x` in the UK but nobody talks about this.
Reposted by Stefan Mayer
Oh yes, Homer’s famous “Tes Thdpssssps.”
👍🏻
👍🏻
February 20, 2025 at 3:51 AM
Oh yes, Homer’s famous “Tes Thdpssssps.”
👍🏻
👍🏻
Reposted by Stefan Mayer
PEOPLE ASSUME THERE CAN BE NO CAUSATION WITHOUT CORRELATION BUT LIKE IF YOU PRESS THE GAS PEDAL IN YOUR CAR AND YOU GO UPHILL THEN SPEED AND GAS PEDAL ARE UNCORRELATED BECAUSE THE SPEED DOESNT CHANGE RIGHT? ALSO IF YOU CONDITION ON A COLLIDER
November 3, 2024 at 10:35 PM
PEOPLE ASSUME THERE CAN BE NO CAUSATION WITHOUT CORRELATION BUT LIKE IF YOU PRESS THE GAS PEDAL IN YOUR CAR AND YOU GO UPHILL THEN SPEED AND GAS PEDAL ARE UNCORRELATED BECAUSE THE SPEED DOESNT CHANGE RIGHT? ALSO IF YOU CONDITION ON A COLLIDER
Reposted by Stefan Mayer
Does the culture you grow up in shape the way you see the world? In a new Psych Review paper, @chazfirestone.bsky.social & I tackle this centuries-old question using the Müller-Lyer illusion as a case study. Come think through one of history's mysteries with us🧵(1/13):
January 25, 2025 at 10:05 PM
Does the culture you grow up in shape the way you see the world? In a new Psych Review paper, @chazfirestone.bsky.social & I tackle this centuries-old question using the Müller-Lyer illusion as a case study. Come think through one of history's mysteries with us🧵(1/13):
Reposted by Stefan Mayer
Yes, they can hallucinate papers that don't exist, discuss results that seem to be imaginary, and can be confusing and inconsistent. But talking to tenured professors may still be helpful
January 14, 2025 at 10:30 PM
Yes, they can hallucinate papers that don't exist, discuss results that seem to be imaginary, and can be confusing and inconsistent. But talking to tenured professors may still be helpful
Reposted by Stefan Mayer
New stroop task just dropped
Another one that I show to demonstrate the perils of improper color usage.
November 20, 2024 at 3:49 PM
New stroop task just dropped
Reposted by Stefan Mayer
I teach a class on data science here at Illinois and during the seminar I host a competition for the class called "The Hunt for the Worst Data Visualization."
This was my most popular post on Xitter, so reproducing some of the best submissions from previous years here:
This was my most popular post on Xitter, so reproducing some of the best submissions from previous years here:
November 11, 2024 at 7:02 PM
I teach a class on data science here at Illinois and during the seminar I host a competition for the class called "The Hunt for the Worst Data Visualization."
This was my most popular post on Xitter, so reproducing some of the best submissions from previous years here:
This was my most popular post on Xitter, so reproducing some of the best submissions from previous years here:
Reposted by Stefan Mayer
a database of datasets ⬇️
been working for awhile on pulling together economic datasets for econ & policy students b/c there is *so* much publicly available data
sharing it for anyone who finds it useful
pls suggest additions; work in progress
lenorepalladino.notion.site/Economic-and...
been working for awhile on pulling together economic datasets for econ & policy students b/c there is *so* much publicly available data
sharing it for anyone who finds it useful
pls suggest additions; work in progress
lenorepalladino.notion.site/Economic-and...
October 16, 2023 at 7:36 PM
a database of datasets ⬇️
been working for awhile on pulling together economic datasets for econ & policy students b/c there is *so* much publicly available data
sharing it for anyone who finds it useful
pls suggest additions; work in progress
lenorepalladino.notion.site/Economic-and...
been working for awhile on pulling together economic datasets for econ & policy students b/c there is *so* much publicly available data
sharing it for anyone who finds it useful
pls suggest additions; work in progress
lenorepalladino.notion.site/Economic-and...
Reposted by Stefan Mayer
Microsoft Word developers. ‘No-one wants to write comments on footnotes.’
PhD Supervisors. ‘?????’
Microsoft Word developers. ‘No-one WHO MATTERS wants to write comments on footnotes.’
PhD Supervisors. ‘?????’
Microsoft Word developers. ‘No-one WHO MATTERS wants to write comments on footnotes.’
April 3, 2024 at 6:39 AM
Microsoft Word developers. ‘No-one wants to write comments on footnotes.’
PhD Supervisors. ‘?????’
Microsoft Word developers. ‘No-one WHO MATTERS wants to write comments on footnotes.’
PhD Supervisors. ‘?????’
Microsoft Word developers. ‘No-one WHO MATTERS wants to write comments on footnotes.’
Reposted by Stefan Mayer
Exactly this 👇
LLM are exactly what the name Large Language Models means. They model language. They construct sentences that sound plausible, even probable. They are not reproducing facts, but rather language that reads like facts. What makes them dangerous is that people don’t understand that.
LLM are exactly what the name Large Language Models means. They model language. They construct sentences that sound plausible, even probable. They are not reproducing facts, but rather language that reads like facts. What makes them dangerous is that people don’t understand that.
Ppl need to stop referring to AI “hallucinations.” LLMs are maybe-most-likely-machines not tools for accurate summarization. It is HIGHLY LIKELY that I would have written an article titled “Manuscript Mediation & Reproduction of Authority” but I didn’t. That’s not hallucination; that’s probability.
February 23, 2024 at 6:41 AM
Exactly this 👇
LLM are exactly what the name Large Language Models means. They model language. They construct sentences that sound plausible, even probable. They are not reproducing facts, but rather language that reads like facts. What makes them dangerous is that people don’t understand that.
LLM are exactly what the name Large Language Models means. They model language. They construct sentences that sound plausible, even probable. They are not reproducing facts, but rather language that reads like facts. What makes them dangerous is that people don’t understand that.
Reposted by Stefan Mayer
Error checking is an important service to science, but it doesn't pay the bills. But what if it did? Introducing ERROR, a bug bounty program for science to systematically detect and report errors in academic publications error.reviews
Home
ERROR is a bug bounty program for science to systematically detect and report errors in academic publications
error.reviews
February 16, 2024 at 5:38 PM
Error checking is an important service to science, but it doesn't pay the bills. But what if it did? Introducing ERROR, a bug bounty program for science to systematically detect and report errors in academic publications error.reviews
More like spilling coffee person ...
February 10, 2024 at 1:09 PM
More like spilling coffee person ...
Reposted by Stefan Mayer
Just realized: one reason the vibes on here are so much better than on Twitter is that the airlines don’t have accounts yet, so my feed lacks those annoying Delta you said I would have access to the Black Diamond VIP room but here I am stuck in the Lapis Lazuli Lounge pls reply immediately posts
December 7, 2023 at 2:44 AM
Just realized: one reason the vibes on here are so much better than on Twitter is that the airlines don’t have accounts yet, so my feed lacks those annoying Delta you said I would have access to the Black Diamond VIP room but here I am stuck in the Lapis Lazuli Lounge pls reply immediately posts
Reposted by Stefan Mayer
Late addition but this easily makes my top 10 in science news this year!
Link: www.surf.nl/en/news/sane...
Link: www.surf.nl/en/news/sane...
❄️The prototype of SANE (Secure Analysis Environment) was launched in 2023!
💻The environment creates a virtual container in which researchers can analyse sensitive data, while the data owner retains full control. The project is a collaborative effort of SURF, CLARIAH, and @odissei.bsky.social
💻The environment creates a virtual container in which researchers can analyse sensitive data, while the data owner retains full control. The project is a collaborative effort of SURF, CLARIAH, and @odissei.bsky.social
December 3, 2023 at 9:46 AM
Late addition but this easily makes my top 10 in science news this year!
Link: www.surf.nl/en/news/sane...
Link: www.surf.nl/en/news/sane...
Reposted by Stefan Mayer
I'm thrilled to announce the new Psych Science editor team! We'll start on Jan 1, 2024. Check us out: here: https://www.psychologicalscience.org/publications/psychological_science/2024-editorial-board
We're spread across 15 countries in 5 continents! 🌍🌎🌏
Here’s a thread to introduce the team!
We're spread across 15 countries in 5 continents! 🌍🌎🌏
Here’s a thread to introduce the team!
www.psychologicalscience.org
December 1, 2023 at 10:04 AM
I'm thrilled to announce the new Psych Science editor team! We'll start on Jan 1, 2024. Check us out: here: https://www.psychologicalscience.org/publications/psychological_science/2024-editorial-board
We're spread across 15 countries in 5 continents! 🌍🌎🌏
Here’s a thread to introduce the team!
We're spread across 15 countries in 5 continents! 🌍🌎🌏
Here’s a thread to introduce the team!
Reposted by Stefan Mayer
Just imagine all the angry e-mails from the people who think the gorilla is missing from their set.
November 20, 2023 at 4:20 PM
Just imagine all the angry e-mails from the people who think the gorilla is missing from their set.
Reposted by Stefan Mayer
Science Magazine coverage: "Preregistering, transparency, and large samples boost psychology studies’ replication rate to nearly 90%"
www.science.org/content/arti...
www.science.org/content/arti...
Our paper "High replicability of newly discovered
social-behavioral findings is achievable" appears today in Nature Human Behaviour.
Just as the large replications provided an inflection point in identifying replicability challenges, this paper does so for solutions.
www.nature.com/articles/s41...
social-behavioral findings is achievable" appears today in Nature Human Behaviour.
Just as the large replications provided an inflection point in identifying replicability challenges, this paper does so for solutions.
www.nature.com/articles/s41...
November 9, 2023 at 8:01 PM
Science Magazine coverage: "Preregistering, transparency, and large samples boost psychology studies’ replication rate to nearly 90%"
www.science.org/content/arti...
www.science.org/content/arti...
Reposted by Stefan Mayer
We are doing some early product testing of algorithm and human assessments of the credibility research claims building on the SCORE project.
If you have recently or will soon preprint social-behavioral empirical research and can give feedback on assessments of your paper please reply or email me!
If you have recently or will soon preprint social-behavioral empirical research and can give feedback on assessments of your paper please reply or email me!
November 4, 2023 at 2:27 PM
We are doing some early product testing of algorithm and human assessments of the credibility research claims building on the SCORE project.
If you have recently or will soon preprint social-behavioral empirical research and can give feedback on assessments of your paper please reply or email me!
If you have recently or will soon preprint social-behavioral empirical research and can give feedback on assessments of your paper please reply or email me!
Reposted by Stefan Mayer
New economics reproducibility project is seeking collaborators to conduct reproduction and robustness tests of published development economics findings.
Contributors earn 2500 euros and co-authorship on the overall project paper.
Details: www.rwi-essen.de/fileadmin/us...
Contributors earn 2500 euros and co-authorship on the overall project paper.
Details: www.rwi-essen.de/fileadmin/us...
November 3, 2023 at 11:11 PM
New economics reproducibility project is seeking collaborators to conduct reproduction and robustness tests of published development economics findings.
Contributors earn 2500 euros and co-authorship on the overall project paper.
Details: www.rwi-essen.de/fileadmin/us...
Contributors earn 2500 euros and co-authorship on the overall project paper.
Details: www.rwi-essen.de/fileadmin/us...
Reposted by Stefan Mayer
Apparently ~40% of crowdsource workers use LLMs when completing surveys
Relatively simple fix for text responses: Add Javascript to Qs to prevent pasting
Credit to Jiabi Wang for idea + Kaushal Addanki for code below
github.com/rafmbatista/...
Relatively simple fix for text responses: Add Javascript to Qs to prevent pasting
Credit to Jiabi Wang for idea + Kaushal Addanki for code below
github.com/rafmbatista/...
October 31, 2023 at 1:52 AM
Apparently ~40% of crowdsource workers use LLMs when completing surveys
Relatively simple fix for text responses: Add Javascript to Qs to prevent pasting
Credit to Jiabi Wang for idea + Kaushal Addanki for code below
github.com/rafmbatista/...
Relatively simple fix for text responses: Add Javascript to Qs to prevent pasting
Credit to Jiabi Wang for idea + Kaushal Addanki for code below
github.com/rafmbatista/...